1. Field of the Invention
The present invention relates to a power-supply-noise cancelling circuit and a solid-state imaging device, and more particularly, is suitably applied to a method of removing power supply noise superimposed on a pixel output signal of a CMOS image sensor.
2. Description of the Related Art
In a CMOS image sensor, a signal from each pixel is sent to a sample-hold signal conversion circuit through a vertical signal line, and a signal component of the signal is detected by CDS (Correlated Double Sampling).
Assuming that a reset-level sampling voltage is Vc and a read-level sampling voltage is Vs, the signal component detected by CDS is given by Vc-Vs.
Meanwhile, when power supply noise is superimposed on the signal from each pixel, the signal component detected by CDS is given by the following Equation (1):(Vc+ΔVc)−(Vs+ΔVs)=(Vc−Vs)+(ΔVc−ΔVs)  (1)where ΔVc is power supply noise at the time of sampling of the reset level and ΔVs is power supply noise at the time of sampling of the read level.
In this case, when the power supply noise ΔVc and the power supply noise ΔVs are equal to each other, the power supply noise ΔVc and the power supply noise ΔVs cancel out each other and degradation of image quality because of the power supply noise ΔVc and the power supply noise ΔVs does not occur. However, because the power supply noise ΔVc and the power supply noise ΔVs are generated randomly, values of the power supply noise ΔVc and the power supply noise ΔVs are generally different from each other.
A method of reducing noise is disclosed in, for example, Japanese Patent Application Laid-open No. 2008-11284, in which an averaging process is performed on a plurality of digital code values obtained by repeating AD conversion multiple times on a difference between potentials at the time of reset of an imaging device and after exposure of the imaging device.
However, in the method disclosed in Japanese Patent Application Laid-open No. 2008-11284, because the averaging process is performed on the difference between the potentials at the time of reset of the imaging device and after the exposure of the imaging device to reduce the noise, an actual signal component at the time of sampling of the read level may not be extracted and resolution may be degraded accordingly. Furthermore, because a signal from an identical pixel needs to be sampled multiple times, processing time may be increased.